Genomic analysis has revealed significant molecular heterogeneity within single tumors. This heterogeneity remains a major obstacle to understanding the biological drivers of tumor progression. To that end, we have developed a new imaging-based technique to obtain genomic profiles of any cell or population by precisely selecting and extracting living cells of interest from their native environment. These purified and extracted cells can then be subjected to genomic analysis, or amplified and cultured for molecular studies. We have termed this image-guided genomics technique, spatiotemporal genomic analysis (SAGA). Due to our long-standing interest in cancer invasion and metastasis, we have used SAGA to probe cancer invasion by focusing on collective tumor cell invasion, which is major mode of metastasis in murine models and patients. We have used SAGA to: 1) perform the first genomic analysis directly comparing the genomic expression profile of purified highly invasive leader cells of the collective invasion pack, to follower cells, which stream behin the leader cells. 2) select, amplify, and maintain the first purified leader cell lines and followe cell lines, which now give us virtually unlimited quantities of previously rare invading cell types Our genomics data show that leader cells have distinct expression profiles with overexpression of key invasion-related genes compared to follower cells. Since we can keep these rare cell types in culture, and they maintain their respective phenotype over time, we show that leader cells invade aggressively yet proliferate poorly, whereas followers show the converse by invading poorly and proliferating rapidly. Importantly, mixing these two cell types leads to a synergy where leader cells reprise their roles and actively seek out follower cells to promote collective invasion; follower cells, in turn, promote effective leader cell proliferation. These observations lead us to tumor cell specialization within the collective invasion pack results in increased metastatic success due to cell-cell cooperativity. hypothesize that If true, it suggests that collectively invading cells represent a specialized labor force that are specifically primed t promote invasion To test this, we will determine if cell specialization within the invasion unit dictates invasive potential, primary tumor formation, and metastasis to a secondary site by moving to in vivo assays with our unique cell lines. Then we will probe the molecular basis of leader cell invasive behavior by leveraging our current genomic data set and determine which genes drive leader and follower cell behavior and communication during collective cell invasion. Taken together, these approaches provide a unique opportunity to understand the mechanistic underpinnings of rare cell types and heterogeneity within the collective invasion pack and proliferation.